Recent years have witnessed considerable interest in (dynamic) issue ownership. While new insights have been gained, progress is stifled by two factors. One, research on issue ownership is typically subject to data sparsity, which has often restricted analyses to few issues. Two, research has mostly studied issue ownership as simple percentages, which are prone to random sampling error, thus disregarding uncertainty in estimating public attributions of issue ownership. To overcome both shortcomings, we propose a Bayesian multilevel model. The model can be flexibly specified to recover dynamic issue ownership. The model is applied to data from the German Longitudinal Election Study. Substantively, the model shows that parties’ issue competences display some malleability, but that changes unfold gradually over time.
